82
Infomation Sleep Deprivation and Cognitive Testing in Internal Medicine House Staff JOHN ROBBINS, MD, and FRED GOTTLIEB, MD, Sacramento, California
There is increased concern about the effects of sleep deprivation on physician performance. We administered four standard tests of cognitive function to 23 university hospital house staff. Each physician served as his or her own control, and the tests were administered at rest, after a night on call, and after a night of sleep for recovery. The study was designed so that normal learning would minimize any deterioration in the post-on-call test performance. Statistically significant deterioration occurred in 3 of the 4 tests after a night on call. Even physicians acclimated to sleep deprivation on a regular, every-third-or-fourth-night basis showed functional impairment. The results have implications for patient care under conditions where house staff are stressed by sleep deprivation and prolonged fatigue. (Robbins J, Gottlieb F: Sleep deprivation and cognitive testing in internal medicine house staff. West J Med 1990 Jan;
152:82-86)
Physicians in residency training programs are required to work for prolonged periods under conditions of sleep deprivation. It is at times of most severe sleep deprivation, that is, "on-call" nights, that residents are most likely to be responsible for the care of acutely ill patients. The therapeutic interventions required at such periods necessitate a high level of psychomotor performance, cognitive function, and emotional equilibrium. The greatest demand on a physician s skills often comes when performance is most likely imiipaired by sleep loss. Despite the large amount of research in the field of sleep deprivation, the data regarding physician performance under sleep-deprived conditions are sparse. Existent information implies potentially adverse effects on mood and behavior. 1-8 Although there are notable exceptions,4'69 few studies specifically assess the psychomotor and cognitive performance of house staff under conditions of acute sleep deprivation. The classic medical study was published almost 20 years ago when Friedman and co-workers showed that sleep-deprived house staff made more errors when interpreting electrocardiograms.9 That study addressed one important cognitive aspect of medical practice. There may be more global deterioration of post-on-call house-staff performance. Deaconson and associates were unable to show the effect of sleep deprivation on surgical residents' performance on psychometric tests. '0 That study used residents on an everyother-night call schedule as their own controls. As pointed out in numerous letters to the editor that followed the original publication, residents in the "control state" may also have been considerably impaired from chronic sleep deprivation. "I To better understand the effects of acute sleep deprivation on house-staff performance, we compared the performance
of house staff on standardized cognitive tests when rested, after being on call-that is, acute sleep deprivation and a third time after conditions of normal sleep. The standard on-call schedule at the institution is every fourth night. The control, or rested, evaluations all took place when the resident had been able to get a full night's sleep at home before taking the test. The study was designed in such a way that if previous sleep deprivation had an effect on the results, it would mitigate against the finding of a statistically significant outcome. The session format allowed us to measure baseline functioning and functioning after sleep deprivation and to recheck the baseline, taking into account the normal learning curve.
Methods Standard tests routinely used in cognitive assessment and rehabilitation were selected to evaluate various cognitive functions. They were administered on a microcomputer terminal with the investigator present at every testing period. The four tests were as follows*: * Driver, in which the subject must maneuver a "car" on a moving track of set width and speed, using a joystick.'2 Scoring is based on percentage accuracy and increased by keeping the car at the bottom of the screen-that is, with less lead time before road hazards appear, which minimizes reaction time. The final score is, therefore, expressed as percentage accuracy and total score. Attention span, eye-hand coordination, and reaction time are assessed. * Rapid Number Comparison, in which four numbers of five digits each are flashed on the screen for approximately 0.5 seconds."' Two numbers are identical, and that number *Bonnie Napier, OT, assisted with the computer programs and testing.
Fr,ow the Department of Internal Medicine, Division of General Medicine, University of California, Davis, Medical Center, Sacramento, California. I)r (Gottlieb is currently with the Northwest Community Health Center, Salt Lake City. Rcprint requcsts to John Robbins, MD, Division of General Medicine, 3120 Primary Care Center, 2221 Stockton Blvd, Sacramento, CA 95817.
THE WESTERN JOURNAL OF MEDICINE
-
JANUARY 1990
*
must be accurately recognized and logged into the computer. Visual reaction time and accuracy, rapid numeric discrimination, short-term memory, and attention span are assessed. * Concentration, in which the subject is presented with a field of characters arranged in a random pattern. 14 A character is selected sequentially by the computer as the target. After the target is located, the subject maneuvers a cursor to cover the character. Scoring is based on the speed oflocating the selected character and moving the cursor minus incorrect responses. If a target is not located in a specified time, the subject must start over. Spatial perception, performance under prolonged repetitive conditions, the ability to discern relevant from random information, eye-hand coordination, and the ability to perform under time limitations are assessed. Assistance was given to complete the test to residents who were unable to do this test in the allotted time under post-on-call conditions. Final scoring was affected only in that it allowed test completion for data interpretation. * Smart Shopper Marathon, referred to hereafter as Math Problems, is a test in which a subject is presented five sets of simple, high-school level mathematic problems under time limitations. 15 Scoring is based on the number of correct answers in each problem set. There is a time penalty for incorrect answers. Each set requires a minute to complete, and simple mathematic abilities under time limitations are assessed. The testing period required for these four tests is about 30 to 40 minutes. Each subject who agreed to participate was given a baseline test under conditions of a normal night of sleep for that person. A second test was given under conditions of sleep deprivation. At this second testing period, the volunteering house officer would page the investigator at his or her own convenience before leaving the hospital on completing call night duties (approximately 35 hours from the end of their last normal sleep period). A third and final testing period was given at the convenience of each subject, again under non-sleep-deprived conditions, to determine if change and performance trends under conditions of sleep deprivation would persist after the subjects were rested. Therefore, a change noted in overall performance under conditions of sleep deprivation should revert to baseline if sleep deprivation is the only variable.
83
* 1 152 152..
TABLE 1.-Cognitive Testing of Sleep Deprivation-Correlation of Scores (r) in Repetitive Testing (Reproducibility)
% Accuracy ..... Total score, r.
.....
Before On-Call Night
After On-Call Night
0.9277 0.8641
0.9298 0.9091
After
Recovery 0.8860 0.9495
To assess the reproducibility of results and the validity of these tests, the test with the greatest potential range of scores, Driver, was administered a second time at each testing session to 17 of the 23 subjects who agreed to donate the additional time required. Before doing the test under conditions of sleep deprivation, each subject estimated the amount of sleep obtained on call and the total number of hours since getting a full night's sleep. This is obviously subjective but was intended to make sure none of the subjects had spent two successive nights in a sleep-deprived state, but that all had some degree of sleep deprivation. Baseline assessments were carried out at various times during the normal working day. In analyzing data, each subject served as his or her own control. No incentive was offered for participation or performance in the study.
Subjects Members of the house staff in the internal medicine department at the University of California, Davis, Medical Center (Sacramento, California) were asked to volunteer in a study to determine the effects, if any, that sleep deprivation typical of a night on call might have on cognitive performance. In all, 31 of the house staffcompleted the first battery of tests under conditions of normal sleep. Eight house staff completed the baseline study but were unable to complete the remainder of the study. Of this group, five reported being too tired after being on call and three were not on rotations requiring being on call during the ten-week period of this study. Thus, 23 house staff completed the study. The maleto-female ratio was 16:7, and the distribution of year in training-9 in their first year after graduating medical school, 7 in their second year, 6 in their third year, and 1 in 82
B
A
5,250 755SEM / 156.
67.5
-
65.0 -I 1.2 SEM
65.0 C
62.5
-5,00015 4,750
-
-
4,500-
60.0-
P
Infomation Sleep Deprivation and Cognitive Testing in Internal Medicine House Staff JOHN ROBBINS, MD, and FRED GOTTLIEB, MD, Sacramento, California
There is increased concern about the effects of sleep deprivation on physician performance. We administered four standard tests of cognitive function to 23 university hospital house staff. Each physician served as his or her own control, and the tests were administered at rest, after a night on call, and after a night of sleep for recovery. The study was designed so that normal learning would minimize any deterioration in the post-on-call test performance. Statistically significant deterioration occurred in 3 of the 4 tests after a night on call. Even physicians acclimated to sleep deprivation on a regular, every-third-or-fourth-night basis showed functional impairment. The results have implications for patient care under conditions where house staff are stressed by sleep deprivation and prolonged fatigue. (Robbins J, Gottlieb F: Sleep deprivation and cognitive testing in internal medicine house staff. West J Med 1990 Jan;
152:82-86)
Physicians in residency training programs are required to work for prolonged periods under conditions of sleep deprivation. It is at times of most severe sleep deprivation, that is, "on-call" nights, that residents are most likely to be responsible for the care of acutely ill patients. The therapeutic interventions required at such periods necessitate a high level of psychomotor performance, cognitive function, and emotional equilibrium. The greatest demand on a physician s skills often comes when performance is most likely imiipaired by sleep loss. Despite the large amount of research in the field of sleep deprivation, the data regarding physician performance under sleep-deprived conditions are sparse. Existent information implies potentially adverse effects on mood and behavior. 1-8 Although there are notable exceptions,4'69 few studies specifically assess the psychomotor and cognitive performance of house staff under conditions of acute sleep deprivation. The classic medical study was published almost 20 years ago when Friedman and co-workers showed that sleep-deprived house staff made more errors when interpreting electrocardiograms.9 That study addressed one important cognitive aspect of medical practice. There may be more global deterioration of post-on-call house-staff performance. Deaconson and associates were unable to show the effect of sleep deprivation on surgical residents' performance on psychometric tests. '0 That study used residents on an everyother-night call schedule as their own controls. As pointed out in numerous letters to the editor that followed the original publication, residents in the "control state" may also have been considerably impaired from chronic sleep deprivation. "I To better understand the effects of acute sleep deprivation on house-staff performance, we compared the performance
of house staff on standardized cognitive tests when rested, after being on call-that is, acute sleep deprivation and a third time after conditions of normal sleep. The standard on-call schedule at the institution is every fourth night. The control, or rested, evaluations all took place when the resident had been able to get a full night's sleep at home before taking the test. The study was designed in such a way that if previous sleep deprivation had an effect on the results, it would mitigate against the finding of a statistically significant outcome. The session format allowed us to measure baseline functioning and functioning after sleep deprivation and to recheck the baseline, taking into account the normal learning curve.
Methods Standard tests routinely used in cognitive assessment and rehabilitation were selected to evaluate various cognitive functions. They were administered on a microcomputer terminal with the investigator present at every testing period. The four tests were as follows*: * Driver, in which the subject must maneuver a "car" on a moving track of set width and speed, using a joystick.'2 Scoring is based on percentage accuracy and increased by keeping the car at the bottom of the screen-that is, with less lead time before road hazards appear, which minimizes reaction time. The final score is, therefore, expressed as percentage accuracy and total score. Attention span, eye-hand coordination, and reaction time are assessed. * Rapid Number Comparison, in which four numbers of five digits each are flashed on the screen for approximately 0.5 seconds."' Two numbers are identical, and that number *Bonnie Napier, OT, assisted with the computer programs and testing.
Fr,ow the Department of Internal Medicine, Division of General Medicine, University of California, Davis, Medical Center, Sacramento, California. I)r (Gottlieb is currently with the Northwest Community Health Center, Salt Lake City. Rcprint requcsts to John Robbins, MD, Division of General Medicine, 3120 Primary Care Center, 2221 Stockton Blvd, Sacramento, CA 95817.
THE WESTERN JOURNAL OF MEDICINE
-
JANUARY 1990
*
must be accurately recognized and logged into the computer. Visual reaction time and accuracy, rapid numeric discrimination, short-term memory, and attention span are assessed. * Concentration, in which the subject is presented with a field of characters arranged in a random pattern. 14 A character is selected sequentially by the computer as the target. After the target is located, the subject maneuvers a cursor to cover the character. Scoring is based on the speed oflocating the selected character and moving the cursor minus incorrect responses. If a target is not located in a specified time, the subject must start over. Spatial perception, performance under prolonged repetitive conditions, the ability to discern relevant from random information, eye-hand coordination, and the ability to perform under time limitations are assessed. Assistance was given to complete the test to residents who were unable to do this test in the allotted time under post-on-call conditions. Final scoring was affected only in that it allowed test completion for data interpretation. * Smart Shopper Marathon, referred to hereafter as Math Problems, is a test in which a subject is presented five sets of simple, high-school level mathematic problems under time limitations. 15 Scoring is based on the number of correct answers in each problem set. There is a time penalty for incorrect answers. Each set requires a minute to complete, and simple mathematic abilities under time limitations are assessed. The testing period required for these four tests is about 30 to 40 minutes. Each subject who agreed to participate was given a baseline test under conditions of a normal night of sleep for that person. A second test was given under conditions of sleep deprivation. At this second testing period, the volunteering house officer would page the investigator at his or her own convenience before leaving the hospital on completing call night duties (approximately 35 hours from the end of their last normal sleep period). A third and final testing period was given at the convenience of each subject, again under non-sleep-deprived conditions, to determine if change and performance trends under conditions of sleep deprivation would persist after the subjects were rested. Therefore, a change noted in overall performance under conditions of sleep deprivation should revert to baseline if sleep deprivation is the only variable.
83
* 1 152 152..
TABLE 1.-Cognitive Testing of Sleep Deprivation-Correlation of Scores (r) in Repetitive Testing (Reproducibility)
% Accuracy ..... Total score, r.
.....
Before On-Call Night
After On-Call Night
0.9277 0.8641
0.9298 0.9091
After
Recovery 0.8860 0.9495
To assess the reproducibility of results and the validity of these tests, the test with the greatest potential range of scores, Driver, was administered a second time at each testing session to 17 of the 23 subjects who agreed to donate the additional time required. Before doing the test under conditions of sleep deprivation, each subject estimated the amount of sleep obtained on call and the total number of hours since getting a full night's sleep. This is obviously subjective but was intended to make sure none of the subjects had spent two successive nights in a sleep-deprived state, but that all had some degree of sleep deprivation. Baseline assessments were carried out at various times during the normal working day. In analyzing data, each subject served as his or her own control. No incentive was offered for participation or performance in the study.
Subjects Members of the house staff in the internal medicine department at the University of California, Davis, Medical Center (Sacramento, California) were asked to volunteer in a study to determine the effects, if any, that sleep deprivation typical of a night on call might have on cognitive performance. In all, 31 of the house staffcompleted the first battery of tests under conditions of normal sleep. Eight house staff completed the baseline study but were unable to complete the remainder of the study. Of this group, five reported being too tired after being on call and three were not on rotations requiring being on call during the ten-week period of this study. Thus, 23 house staff completed the study. The maleto-female ratio was 16:7, and the distribution of year in training-9 in their first year after graduating medical school, 7 in their second year, 6 in their third year, and 1 in 82
B
A
5,250 755SEM / 156.
67.5
-
65.0 -I 1.2 SEM
65.0 C
62.5
-5,00015 4,750
-
-
4,500-
60.0-
P
